Elevational diversity of terrestrial rainforest herbs:
when the whole is less than the sum of its parts
Sandra Willingho¨fer• Daniele Cicuzza•Michael Kessler
Received: 23 January 2011 / Accepted: 29 September 2011 / Published online: 12 October 2011 Ó Springer Science+Business Media B.V. 2011
Abstract We studied the species richness of herba-ceous terrestrial plant species along an elevational gradient at 250–2425 m a.s.l. in evergreen tropical forest in Central Sulawesi, Indonesia. We recorded 302 species belonging to 51 families. Ferns and lycophytes contributed 62% of the species, followed by monocots with 24% and dicots with 14%. Overall herb species richness did not show any particular relation with elevation, while the richness of ferns increased significantly with elevation, monocots did not show a pattern, and dicots showed a hump-shaped pattern with maximum richness at 1800 m. These patterns in turn were only partly reflected in the patterns of the individual plant families making up each group. The independence of different taxa was also reflected in their relationships to environmental factors (temperature, precipitation, and area): although, each single family was related to one or several factors, at the group level and at the overall
level these trends were lost. These results show that interpreting diversity at higher taxonomic level may overlook important information at the family level and raises the biologically intriguing question whether overall patterns of diversity result from a random accumulation of group-specific patterns or if there is some interaction between groups (e.g., via competi-tion and niche-pre-empcompeti-tion).
Keywords Elevational gradient Terrestrial herbs Tropics Diversity Species richness Sulawesi Indonesia
Introduction
Tropical rainforests are well known for their enormous plant and animal diversity, which is especially concentrated in mountain regions (Myers et al.2000; Kreft and Jetz2007). While the number of field studies on tropical plant diversity is constantly increasing, this research effort is taxonomically and ecologically unevenly distributed, focussing on trees at the expense of other life forms (lianas, shrubs, terrestrial, and epiphytic herbs) (Linares-Palomino et al. 2009). Although, terrestrial herbs are recognized as an important component in tropical plant communities (Gentry 1995; Gentry and Dodson 1987; Linares-Palomino et al. 2009), there are few quantitative studies of the diversity of tropical herbs. The terrestrial Electronic supplementary material The online version of
this article (doi:10.1007/s11258-011-9986-z) contains supplementary material, which is available to authorized users. S. Willingho¨fer
Department of Systematic Botany, Albrecht von Haller Institute of Plant Sciences, Untere Karspu¨le 2, 37073 Go¨ttingen, Germany
D. Cicuzza (&) M. Kessler
Institute of Systematic Botany, University of Zu¨rich, Zollikerstrasse 107, 8008 Zurich, Switzerland e-mail: daniele.cicuzza@systbot.uzh.ch DOI 10.1007/s11258-011-9986-z
layer of herbaceous plant species in tropical forests plays an important role in light limitation and space competition for seedling and juvenile trees (Tsvuura et al. 2010), influencing tree regeneration and forest dynamics (Condit et al.2000), but there is very little data on the diversity of terrestrial herb assemblages (e.g., Poulsen and Pendry 1995). Along tropical elevational gradients, for example, there are dozens of studies on the diversity of trees (e.g., Kitayama
1992) and a lesser number on lianas (Homeier et al.
2010), ferns (Kluge et al.2006; Kessler2000, Kessler et al. 2011) and epiphytes (Kro¨mer et al. 2005; Cardelu´s et al. 2006), but we are aware of only a handful of studies including terrestrial herbs, although none of these specifically focuses on herbs. Moreover, these studies have all recovered different elevational species richness patterns: Poulsen and Pendry (1995) documented a U-shaped pattern along a short eleva-tional gradient (200–850 m) in Borneo, Poulsen et al. (2005) a monotonic decline between 770 and 1520 m in central Africa, Grytnes and Beaman (2006) a hump-shaped pattern along an extensive gradient (0–4094 m) in Borneo, and Desalegn and Beierkuhn-lein (2010) a roughly monotonic increase with eleva-tion in Ethiopia (1200–2700 m). Thus, little is known on how the diversity and community composition of forest herbs change with elevation, whether there is any generality to them, and if these patterns corre-spond with those of other life forms.
Elevational gradients offer an outstanding oppor-tunity to study and understand patterns of biotic diversity because they represent reproducible envi-ronmental gradients that can be replicated across the globe (Lomolino 2001; McCain 2009). Approxi-mately, half of all case studies across a wide range of taxa show hump-shaped richness patterns with maximum richness at an intermediate point of the gradient, with the remainder of the studies including monotonic increases or decreases as well as roughly constant numbers (Rahbek1995,2005; McCain2007,
2009). The causes for these different patterns are largely unknown but are often taxon-specific and depend on the spatial extent as well as the geograph-ical location of the gradients (Rahbek2005; Kessler et al.2011).
A large number of partly non-exclusive explana-tions have been proposed for taxon-specific elevation-al richness patterns, including (1) current climatic variables such as temperature and humidity (Heaney
2001; Kessler 2001; Bhattarai et al. 2004) which in turn determine energy availability and ecosystem productivity (Hawkins et al. 2003; Currie et al.
2004), (2) spatial aspects including regional area size (Rosenzweig and Ziv1999) and geometric constraints (Colwell et al.2004; Grytnes et al.2008), (3) historical and evolutionary processes (Ricklefs 2004), and (4) biotic processes such as the Rapoport rescue hypoth-esis (Stevens 1989) or source-sink effects (Grytnes
2003; Kessler2009). All of these factors may covary with both hump-shaped and linear species richness patterns, and discrimination between them is fre-quently difficult, especially due to a lack of relevant ecological data. Despite these limitations, there is increasing consensus that there are close links between species richness and climatic factors, and that factors such as area, geometric constraints and population-level processes at best modify the climate-dependent patterns (Currie and Kerr2008; McCain2009; Kessler et al.2011).
The objective of this study was to document the patterns of species richness of terrestrial forest herbs along an elevational gradient on the island of Sulawesi in Indonesia and to identify possible correlations with environmental factors. Specifically, we asked the following questions: (1) How does terrestrial herba-ceous species richness change along the elevational gradient? (2) Do different taxonomic groups show similar patterns? and (3) Which are the environmental and spatial factors correlated with herbaceous species richness?
Materials and methods
Study sites and field sampling
The study was conducted in and around the Lore Lindu National Park (LLNP) in Central Sulawesi (Indone-sia). The park encompasses an elevational gradient from 200 to 2509 m a.s.l., with about 90% of the area above 1000 m. Precipitation ranges from 2000 to 4000 mm and mean annual temperatures from 26 to 31°C (The Nature Conservancy,2004).
Field sampling was conducted at eight different sites covering elevations between 250 and 2425 m: Saluki, Mount Nokilalaki (subdivided into 4 eleva-tional belts), Pono, Bariri, and Mount Rorekatimbu (Table 1, Fig.1). The highest peak in Central
Sulawesi is Mount Rorekatimbu (2509 m), hence our transect covered almost the entire elevational range. Sampling was conducted in 2006 at Bariri and from February to April 2008 at the remaining sites. We established a set of three study transects at each locality, with the exception of the summit of Mount Nokilalaki, where only a single transect was placed, resulting in a total of 22 transects (Table1). Each transect consisted of 12 plots of 5 m 9 5 m each, separated by intervals of 35 m from each other, resulting in a total transect length of 445 m. The area of each transect was 300 m2for a total area of 6600 m2 for the whole study. Transects were placed in mature forest, avoiding gaps and other anthropogenic and natural disturbances, water-logged soils or open rock faces. Transects were located following the contour lines to avoid variation in elevation within the transect span.
One possible concern for this study is that sampling took place on several mountains, rather than across the full elevational gradient of a single mountain. This
was necessary because of high rates of deforestation in the study region and limited accessibility due to the steep terrain. However, the three transects conducted at the two localities (Bariri and Nokilalaki) at 1400 m have identical species numbers, suggesting that our study sites are regionally representative.
In each study plot, all terrestrial herbaceous species as well as elevation (measured with a GARMIN GPS) and slope were recorded. All plant species were collected, subsequently identified where possible to species level, and finally deposited in the following herbaria: Go¨ttingen (GOET), Bogoriense Bogor (BO), Celebense Palu (CEB), Leiden (L), Zurich (Z), and Berkeley (UC; ferns only).
As environmental variables, we included mean annual temperature, mean annual precipitation, and land surface area. Climatic parameters for each study site were extracted from WorldClim (Hijmans et al.
2006). Area was evaluated by estimating the island surface of Sulawesi subdivided by elevational belts: 0–500, 501–1000, 1001–1500, 1501–2000, 2001–2500, Table 1 Overview of the
study sites, with information on elevation and climate
Study site Coordinates Elevation (m) Mean annual
precipitation (mm) Mean annual temperature (°C) Saluki 1°18025.2100S 119°57058.2000E 250 1890 24.3 Saluki 1°18019.6400S 119°58059.8600E 300 1890 24.3 Saluki 1°17046.6000S 119°58033.2700E 350 1890 24.3 Pono 1°29020.5400S 120°3030.2500E 850 1895 21.3 Pono 1°2906.9400S 120°3017.1100E 900 1895 21.3 Pono 1°29041.7600S 120°3045.2900E 1120 1895 21.3 Nokilalaki 1° 13021.1500S 120° 9018.5400E 1200 1820 19.6 Nokilalaki 1°13031.8800S 120°9016.2100E 1200 1820 19.6 Nokilalaki 1°13041.1400S 120°9015.2400E 1265 1820 19.6 Bariri 1°39048.3400S 120°10015.7800E 1400 1975 19.1 Bariri 1°39038.4800S 120°10011.6900E 1400 1975 19.1 Bariri 1°39028.3500S 120°1002.3200E 1400 1975 19.1 Nokilalaki 1°13080.9500S 120°901.1800E 1400 1840 18.9 Nokilalaki 1°13048.0200S 120°905.0100E 1400 1840 18.9 Nokilalaki 1°1404.1200S 120°9018.5100E 1490 1840 18.9 Nokilalaki 1°14021.8700S 120°9020.8700E 1800 1875 18.3 Nokilalaki 1°14035.4800S 120°8056.2500E 1800 1875 18.3 Nokilalaki 1°14046.6300S 120°9014.7900E 1880 1875 18.3 Nokilalaki 1°14045.3200S 120°9087.2100E 2200 1940 17.0 Rorekatimbu 1°16021.3400S 120°18075.8100E 2350 2132 14.1 Rorekatimbu 1°16042.6600S 120°18034.1400E 2400 2132 14.1 Rorekatimbu 1°16052.4700S 120°18033.4100E 2425 2132 14.1
and 2501–3000 m. We did not analyse variations of species numbers at the level of plots within transects across the elevational gradient because plots within a transect are spatially autocorrelated.
Statistical analyses
For the analyses, we calculated the mean number of species per plot for the following taxonomic groups: all terrestrial herbaceous plant species together,
separated into three major taxonomic groups (ferns and lycophytes, henceforth called ferns; monocots; dicots), and for the three most species rich families of each of the major groups (Table 2). Although, not monophyletic, we treated ferns and lycophytes as a single group due to the low number of lycophytes and the physiological and ecological similarities between the groups. Likewise, dicots were treated as a group because our sample mainly included eudicots and very few basal dicots. In an earlier phase of the analyses, we Fig. 1 Location of the
protected area and the five focus study area, in the small quadrate the map of the Sulawesi Island with the border of the Lore Lindu National Park in central Sulawesi
also calculated total numbers of species per plot as well as estimated total species richness per plot applying the Chao2 richness estimator using Esti-mateS (Colwell2008). As the results of all three levels of species richness were essentially identical, we only report the mean species number per plot.
To assess the elevational richness patterns of each group, simple linear and polynomial regression mod-els were used to correlate the richness with elevation. Next, stepwise multiple linear regression analysis was used to select the combination of environmental variables that was most strongly related to herbaceous species richness. Elevation was not included as a variable in these analyses because it does not directly influence species richness (Ko¨rner2000) and because it is closely correlated to temperature. For model selection, we used the Akaike Information Criterion (AIC), which incorporates the maximized log-likeli-hood of the model and a term that penalizes models
with greater complexity (Johnson and Omland2004). The model with the lowest AIC value is considered as the best fitting model (Johnson and Omland2004), but when the AIC difference between competing models is small, choosing a single ‘‘best model’’ may lead to a faulty interpretation (Wagenmakers and Farrell2004). Models with DAIC \ 2 all have a substantial support, while there is less support when DAIC [ 4 (Burnham and Anderson 2002). Based on DAIC, the Akaike weight quantifies that a given model is better than the next model within the set of all models. Weighting the different models, we estimated the relative variable importance (RVI) across all models by summing the weights of all competing models including this variable (Burnham and Anderson2002), and the sums for the variables can then be compared. All analyses were conducted with R version 2.10 (R Development Core Team2010), with additional functions provided by the R package vegan (Oksanen et al.2007).
Results
In total, we recorded 302 terrestrial herbaceous plant species belonging to 51 families. Of these, 163 were identified to species level (54%), 97 to genus level (32%), and 42 to family level (14%). Ferns and lycophytes contributed 62% of the species and 41% of the families, monocots 24% of the species and 22% of the families, and dicots 14% of the species and 37% of the families. At the family level, the most species rich families were: (1) Polypodiaceae (29 species), Aspleniaceae (28), and Dryopteridaceae (16), for ferns, (2) Araceae (28), Orchidaceae (19), and Zingiberaceae (14) for monocots, and (3) Gesneriaceae (8), Begoniaceae (5) and Urticaceae (5), for dicots.
Among the environmental variables, temperature and elevation (R = -0.98, P \ 0.001) as well the area and elevation (R = -0.88, P \ 0.001) and precipita-tion and elevaprecipita-tion (R = 0.56, P = 0.006) were corre-lated with each other. Area was significantly correlated to temperature (R = 0.85, P \ 0.001) but not to precipitation (R = -0.29, P = 0.195) (Suppl. Mat. Fig. S1).
The mean number of species per plot of all herbaceous taxa combined was, at best, weakly related to elevation (Fig.2). However, when we analyzed the three main subgroups separately, we recovered dis-tinct patterns for each of them. Fern richness increased Table 2 Results of the multiple regression models relating the
species richness of our study groups to three environmental factors
Taxa/Families Area Mean annual temperature Mean annual precipitation Total species 0.82 (*; -) 0.37 0.19 Ferns 1 (**) 0.44 1 (**; ?) Aspleniaceae 1 0.27 – Dryopteridaceae 0.46 0.32 – Polypodiaceae 0.68 0.32 0.2 Monocots 0.3 1 (*; ?) – Araceae 0.3 1 0.42 Orchidaceae – – – Zingiberaceae 0.68 0.32 – Dicots 0.36 0.64 (**) 1 (***; -) Urticaceae 0.22 0.34 0.83 Gesneriaceae 0.28 1 1 Begoniaceae 0.44 0.56 1
Presented is the relative variable importance (RVI) for each environmental factor and the species richness of our study groups. Values close to 1 show that the respective factor played an important role in all competing multiple regression models with AIC values \2 from the best model, values close to 0 that the factor was unimportant in these models. Dashes reveal cases where the factors were not included in the models. For each group we present the best model and the directions of correlation, representing by * the P value of each factor included in the model and (?) positive and (-) the correlation direction
significantly with elevation (R = 0.60, P \ 0.001) while monocots evidenced no clear elevational trend (R = 0.15, P = 177 0.077), and dicots exhibited a hump-shaped pattern with highest richness at about 1800 m (Fig.2). When we further considered the three most species-rich families per major taxonomic group, we also recovered distinct patterns. Among ferns, the species richness of Dryopteridaceae (R = 180 0.28, P = 0.013), Aspleniaceae (R = 0.27, P = 0.013), and Polypodiaceae (R = 0.20, P = 0.036) all increased with elevation (Fig.2), in accordance with the general pattern for all ferns. Among monocots, Araceae (R = 0.07, P \ 0.001) and Orchidaceae (R = 0.71, P = 0.001) showed decreasing richness with elevation, whereas that of Zingiberaceae increased (R = 0.27, P = 0.014) (Fig.2). Finally, among dicots, Urticaceae showed a significant decrease (R = 0.27, P = 0.013), Gesneriaceae a sig-nificant increase (R = 0.55, P \ 0.001) and Begoni-aceae a non-significant pattern (R = 0.05, P = 0.331) (Fig.2).
When we used multiple regression models to assess the explanatory power, using the relative variable importance (RVI) analysis using Akaike weights, the overall herb richness showed twice as high RVI values, for area than for mean annual temperature and four times higher than for mean annual precipitation (Table2; Fig.3). For ferns, area and precipitation had equally high RVI values about double that of temper-ature. At the family level, the species richness for each of the three most important families, Aspleniaceae, Dryopteridaceae, and Polypodiaceae, was related most strongly to area, followed by temperature. Only the model for Polypodiaceae included the precipitation factor, which had a very low RVI value. The overall pattern of species richness for monocots was related to temperature, followed by area; at the family level Araceae were related to temperature followed by precipitation and area, while the richness of Zingiber-aceae species was related to area and temperature. In contrast, Orchidaceae were not related to any of the three factors. The overall dicot pattern showed the highest RVI value for precipitation followed by temperature and area. Urticaceae were supported by precipitation and less by temperature and area, Ges-neriaceae were equally supported by precipitation and temperature, and less by area, in conclusion Begon-iaceae by precipitation followed by temperature and area (Table2, S1).
A major concern is the locality of Pono (1100 m) which has a particularly high diversity of herbaceous species and strongly influences the overall results. Despite noting this high diversity already during field work, we are unable to provide any convincing explanation at this point. Probably, local ecological conditions are involved, since nearby localities had much lower fern diversity (D. Karger and M. Kessler, unpublished data). We refrained from excluding the locality of Pono, however, because it was selected a priori based on the same criteria as the other sites and its exclusion solely on the basis of a lack of fit to the overall pattern would be subjective.
Discussion
For each of our three initial research questions, we obtained distinct answers: although, total terrestrial herbaceous plant richness did not show a significant change with elevation (question 1), when we analysed the major taxonomic groups and families indepen-dently, most of them showed clear and often contrast-ing patterns (2), in many cases related to different sets of environmental factors (3). Or, in other words, the indistinct overall pattern resulted from the accumula-tion of distinct group-specific patterns. The overall lack of a clear elevational richness pattern for terrestrial herbs parallels the results from New Zea-land forests (Ohlemu¨ller and Wilson 2000) but contrasts with other studies, such as a U-shaped pattern found in Borneo (200–850 m) (Poulsen and Pendry 1995), a monotonic decline between 770 and 1520 m in central Africa (Poulsen et al. 2005), a hump-shaped pattern from sea level to 4094 m in Borneo (Grytnes and Beaman 2006), and a roughly monotonic increase between 1200 and 2700 m in Ethiopia (Desalegn and Beierkuhnlein 2010). This complexity may partly be the result of the different elevational spans covered by the studies, but likely also reflects differences in the ecological conditions between the study regions. It may also be illustrative to compare the elevational gradient to the latitudinal pattern, where a recent study has shown that the richness of herbaceous angiosperm families shows a much less pronounced decline away from the equator than the family richness of trees (Hawkins et al.2011). The authors of the latter study suggest that this may be because herbs have shorter generation times than trees
and are morphologically more flexible in their adap-tations to low temperatures, enabling them to more quickly diversify at higher latitudes. In our study region, tree richness declines roughly constantly with elevation (Culmsee and Pitopang2009; Culmsee et al.
2011) whereas herb diversity does not, pointing to a situation similar to that along the latitudinal gradient and suggesting that a similar mechanism may be at play along both geographical gradients.
Yet, while there was no clear elevational richness pattern for herbs overall, the major groups and families did show distinct patterns, which raises two interre-lated follow-up questions. First, what are the causes of these group-specific patterns? And second, is the fact that the addition of group-specific patterns results in an indistinct overall pattern a chance event, or do the groups interact in some way so that overall herb species richness is restricted to a certain upper bound? Fig. 2 Regression
relationships between the mean number of species per plot to elevation for all terrestrial herb species (top), the three major taxonomic groups (2ndrow) and the three most species rich families in each of these (bottom three rows)
Focusing first on the individual groups, terrestrial fern species richness increased with elevation overall as well as among the three most species rich families (Fig.2). This is a typical pattern in tropical mountains, where species richness of ferns peaks between 1500 and 2500 m, as shown by a worldwide comparison of elevational fern transects (Kessler et al. 2011). Monocots showed no clear pattern overall, as a result
of the addition of family-specific patterns of a decrease in Araceae and Orchidaceae, and an increase in Zingiberaceae. Araceae are well known to be best represented in humid lowland tropical forests and to decrease in their richness with decreasing tempera-tures (Kessler2001; Mayo et al.1997). In the central Himalayas, regional richness of terrestrial orchids increases from the lowlands to a maximum at about Fig. 3 Visualization of the
explanatory power of the three environmental variables (mean annual temperature, mean annual precipitation, and area) on the species richness of all terrestrial herb species (top), the three major taxonomic groups (2ndrow), and the
three most species rich families in each of these (bottom three rows). The sizes of the boxes are proportional to the percentage of variance explained by each factors, with overlapping areas showing variation that cannot be assigned to an individual factor
2000 m (Acharya et al.2011), in parallel to our results at the local scale. These results are supported by our multiple regression model for the Araceae, which includes temperature as a significant factor, but not for the Orchidaceae (Table2). Unfortunately, little is known about elevational richness patterns of Zingi-beraceae. The dicots showed a hump-shaped pattern overall, resulting from a combination of decreasing (Urticaceae), increasing (Gesneriaceae), and indistinct (Begoniaceae) patterns. No comparable studies exist for any of these groups so that the generality of the patterns observed in this study cannot be assessed. However, we found patterns consistent with those groups for which data from previous studies are available (ferns, aroids, and orchids), which suggests that findings from our study may be representative for other groups as well.
In the regression models, the overall dicot richness as well as that of Gesneriaceae and Begoniaceae decreased with temperature (Table2). Precipitation had a negative relationship with the richness of all dicots, Begoniaceae and Urticaceae, whereas the richness of Gesneriaceae increased with the precipi-tation. Because of the paucity of studies on tropical forests herbs, we are unable to assess if these patterns are typical for these families and if they are represen-tative for tropical environments. It is noteworthy, however, that several of our models included a negative relationship between species richness to area. While this result has statistical significance, it is biologically meaningless since there are no conceiv-able mechanisms by which increasing land surface area could reduce the species richness of a group. Most likely, this is a spurious result in which area captured part of the variation of species richness that is in effect accounted for by some ecological factor not measured by us (e.g., soil parameters, air humidity).
As detailed in the introduction, the covariance of explanatory factors along elevational gradients frequently makes it difficult to identify causal rela-tionships. Despite these limitations, there are three non-exclusive major processes that may lead to the observed climate–species richness relationships.
First, the relationship may be determined by the physiological tolerances of the species involved. For example, high temperatures and low air humidity have repeatedly been proposed as limiting fern diversity in lowland tropical forests (Bhattarai et al.2004; Kessler
2001; Kluge et al. 2006), and there is increasing
evidence that ferns are less capable of optimizing their water-use efficiency than angiosperms (Brodribb and Holbrook 2004; Brodribb et al. 2009). Furthermore, the majority of modern ferns have very sensitive photoreceptors (Doi and Shimazaki 2008). This appears to limit the productivity and competitive ability of many ferns and perhaps concentrates them in more shady and humid environments such as mossy mountain forests. Globally, diversity of ferns declines more strongly toward arid and cold climatic conditions than does that of angiosperms (Kreft et al. 2010), further suggesting that ferns are physiologically less adaptable. Similar arguments could be made for each of the other taxa, but because the physiologically most suitable climatic conditions are often deduced from where species richness is highest, there is always a risk of circular argumentation.
Second, there may also be an indirect effect of climate via ecosystem productivity (Lieth 1975; Currie et al. 2004) and hence the energy available for plant growth and development. Evans et al. (2005) proposed a number of mechanisms by which higher productivity can lead to higher species richness, e.g., because species are able to use more narrowly delimited niches. Unfortunately, we are currently unable to test these hypotheses because of the close covariance of productivity to climate, and because we lack productivity measures that are independent of climatic data. Furthermore, we do not know what proportion of the overall ecosystem productivity is taken up by the terrestrial herb assemblages and the individual groups at a given site and elevation.
Third, the high species richness of different groups at different elevations may be an evolutionary legacy of longer lineage ages, or of higher speciation or lower extinction rates under certain climatic conditions (Jablonski et al. 2006; Kozak and Wiens 2010), as has been suggested for the latitudinal pattern of the richness of herbaceous angiosperm families (Hawkins et al. 2011). Testing these hypotheses will require highly resolved and dated phylogenies, which are currently unavailable for our study groups and are generally limited in regions where plant taxonomies are poorly known, such as Sulawesi.
These various processes may have different impacts at different taxonomic levels, which bring us to the question of the causes. Ill-defined overall elevational richness patterns of plants resulting from the addition of distinct group-specific patterns have
previously been documented by Bhattarai and Vetaas (2003) in Nepal and Desalegn and Beierkuhnlein (2010) in Ethiopia, suggesting that the pattern observed in this study may be common in tropical mountains. However, Sa´nchez-Gonza´lez and Lo´pez-Mata (2005) found congruent hump-shaped richness patterns for all plants and major groups in Mexico.
We can conceive two fundamentally different ways in which group-specific richness patterns may be added up. First, the individual groups may be independent of each other so that the indistinct overall pattern may simply be result of the random addition of the group-specific patterns. Second, overall herb species richness at a given elevation may be limited by space or resource availability (Bartels and Chen
2010). In the case that such an upper bound exists, species richness of a given lower-level taxonomic group would not be independent from the diversity of other groups. These hypotheses could be tested experimentally by, e.g., partly removing the herb vegetation from selected plots and observing the possible invasion by species previously absent. Another option would be to compare biogeographi-cally disparate elevational transects along which the composition of, e.g. monocot families, varies, so that the intensity of interfamilial competition would shift. Clearly, species richness patterns of tropical herb species have been largely neglected so far but, as shown in this study, represent a potentially highly interesting field of research that may complement the extensive research conducted especially on trees.
Acknowledgments We thank W. Lorenz, M. Sigit, and A. Rahman for their coordination in Palu and Go¨ttingen, and the local field assistants for their help during field work. We thank Dr. Stefan Erasmi for supporting the cartography of the Lore Lindu National Park, and we thank the two anonymous reviewers for their constructive comments on the earlier version of the manuscript. Field work was supported by the STORMA project (Stability of Tropical Rainforest Margins) funded by the German Research Foundation (DFG). Identification of herbaceous species were conducted partly at the Herbarium of Leiden supported by the European project SYNTHESYS (NL-TAF-3536).
References
Acharya KP, Vetaas OR, Birks HJB (2011) Orchid species richness along Himalayan elevational gradients. J Bioge-ogr.http://wileyonlinelibrary.com/journal/jbi,doi:10.1111/ j.1365-2699.2011.02511.x
Bartels SF, Chen HYH (2010) Is understory plant diversity driven by resource quality or resource heterogeneity? Ecology 91:1931–1938
Bhattarai KR, Vetaas OR (2003) Variation in plant species richness of different life forms along a subtropical eleva-tion gradient in the Himalayas, east Nepal. Global Ecol Biogeogr 12:327–340
Bhattarai KR, Vetaas OR, Grytnes JA (2004) Fern species richness along a central Himalayan elevational gradient, Nepal. J Biogeogr 3:389–400
Brodribb TJ, Holbrook NM (2004) Stomatal protection against hydraulic failure: a comparison of coexisting ferns and angiosperms. New Phytol 162:663–670
Brodribb TJ, McAdam SAM, Jordan GJ, Field TS (2009) Evaluation of stomatal responsiveness to CO2and
opti-mization of water-use efficiency among land plants. New Phytol 183:839–847
Burnham KP, Anderson DR (2002) Model selection and mul-timodel inference: a practical information-theoretic approach. Springer, New York
Cardelu´s LC, Colwell RK, Watkins JEJ (2006) Vascular epi-phyte distribution patterns: explaining the mid-elevation richness peak. J Ecol 94:144–156
Colwell RK (2008) EstimateS: statistical estimation of species richness and shared species from samples. Version 8.0.0 User’s guide and application. Available at http://purl. oclc.org/estimates
Colwell RK, Rahbek C, Gotelli NJ (2004) The mid-domain effect and species richness patterns: what have we learned so far? American Naturalist 163:1–23
Condit R, Ashton PS, Baker P, Bunyavejchewin S, Gunatilleke S, Gunatilleke N, Hubbell SP, Foster RB, Itoh A, LaFrankie JV, Lee HS, Losos E, Manokaran N, Sukumar R, Yamakura T (2000) Spatial patterns in the distribution of tropical tree species. Science 288:1414–1418
Culmsee H, Pitopang R (2009) Tree diversity in sub-montane and lower montane primary rainforests in Central Sulaw-esi. Blumea 54:119–123
Culmsee H, Ramadhanil P, Mangopo H, Sabir S (2011) Tree diversity and phytogeographical patterns of tropical high mountain rain forests in Central Sulawesi, Indonesia. Biodivers Conserv 20:1103–1123
Currie DJ, Kerr JT (2008) Tests of the mid-domain hypothesis: a review of the evidence. Ecol Monogr 78:3–18
Currie DJ, Mittelbach GG, Cornell HV, Field R, Gue´gan JF, Hawkins BA, Kaufman DM, Kerr JT, Oberdorff T, O’Brien E, Turner JRG (2004) Predictions and tests of climate-based hypotheses of broad-scale variation in tax-onomic richness. Ecol Lett 7:1121–1134
Desalegn W, Beierkuhnlein C (2010) Plant species and growth form richness along altitudinal gradients in the southwest Ethiopian highlands. J Veg Sci 21:617–626
Doi M, Shimazaki K (2008) The stomata of the fern Adiantum capillus-veneris do not respond to CO2in the dark and open
by photosynthesis in the guard cells. Plant Physiol 147:922–930
Evans KL, Warren PH, Gaston KJ (2005) Species-energy rela-tionships at the macroecological scale: a review of the mechanisms. Biol Rev 80:1–25
Gentry AH (1995) Patterns of diversity and floristic composition in neotropical montane forests. In: Churchill SP, Balslev H,
Forero E, Luteyn JL (eds) Biodiversity and conservation of neotropical montane forests. The New York Botanical Garden, New York, pp 103–126
Gentry AH, Dodson C (1987) Contribution of nontrees to spe-cies richness of a tropical rain forest. Biotropica 19:149–156
Grytnes JA (2003) Species richness patterns of vascular plants along seven altitudinal transects in Norway. Ecography 26:291–300
Grytnes JA, Beaman JH (2006) Elevational species richness patterns for vascular plants on Mount Kinabalu, Borneo. J Biogeogr 33:1838–1849
Grytnes JA, Beaman JH, Romdal TS, Rahbek C (2008) The mid-domain effect matters: simulation analyses of range-size distribution data on Mount Kinabalu, Borneo. J Biogeogr 35:2138–2147
Hawkins BA, Field R, Cornell HV, Currie DJ, Gue`gan JF, Kaufman DM, Kerr JT, Mittelbach GG, Oberdorff T, O’Brien EM, Porter EE, Turner JRG (2003) Energy, water, and broad-scale geographic patterns of species richness. Ecology 84:3105–3117
Hawkins BA, Rodrıguez MA, and Weller SG (2011) Global angiosperm family richness revisited: linking ecology and evolution to climate. J Biogeogr.http://wileyonlinelibrary. com/journal/jbi,doi:10.1111/j.1365-2699.2011.02490.x
Heaney LR (2001) Small mammal diversity along elevational gradients in the Philippines: an assessment of patterns and hypotheses. Glob Ecol Biogeogr 10:15–39
Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2006) The WorldClim interpolated global terrestrial climate surfaces. Version 1.4.http://www.worldclim.org
Homeier J, Engler F, Leuschner C, Weigelt P, Unger M (2010) Factor controlling the abundance of lianas along an alti-tudinal transect of tropical forest in Ecuador. Forest Ecol Manag 259:1399–1405
Jablonski D, Roy K, Valentine JW (2006) Out of the tropics: evolutionary dynamics of the latitudinal diversity gradient. Science 314:102–106
Johnson JB, Omland KS (2004) Model selection in ecology and evolution. Trends Ecol Evolut 19:101–108
Kessler M (2000) Elevational gradients in species richness and endemism of selected plant groups in the central Bolivian Andes. Plant Ecol 149:181–193
Kessler M (2001) Patterns of diversity and range size of selected plant groups along and elevational transect in the Bolivian Andes. Biodivers Conserv 10:1897–1920
Kessler M (2009) The impact of population processes on pat-terns of species richness: Lessons from elevational gradi-ents. Basic Appl Ecol 10:295–299
Kessler M, Kluge J, Hemp A, Ohlemu¨ller R (2011) A global comparative analysis of elevational species richness pat-terns of ferns. Global Ecol Biogeogr. doi: 10.1111/j.1466-8238.2011.00653.x, http://wileyonlinelibrary.com/journal/ geb
Kitayama K (1992) An altitudinal transect study of the vegeta-tion of Mount Kinabalu, Borneo. Vegetatio 102:149–171 Kluge J, Kessler M, Dunn RR (2006) What drives elevational
patterns of diversity? A test of geometric constraints, cli-mate and species pool effects for pteridophytes on an ele-vational gradient in Costa Rica. Global Ecol Biogeogr 15:358–371
Ko¨rner C (2000) Why are there global gradients in species richness? Mountains might hold the answer. Trends Ecol Evol 15:513–514
Kozak KH, Wiens JJ (2010) Accelerated rates of climatic-niche evolution underlies rapid species diversification. Ecol Lett 13:1378–1389
Kreft H, Jetz W (2007) Global patterns and determinants of vascular plant diversity. Proc Natl Acad Sci USA 14:5925–5930
Kreft H, Jetz W, Mutke M, Barthlott W (2010) Contrasting environmental and regional effects on global pteridophyte and seed plant diversity. Ecography 33:408–419 Kro¨mer T, Kessler M, Gradstein SR, Acebey A (2005) Diversity
patterns of vascular epiphytes along an elevational gradient in the Andes. J Biogeogr 32:1799–1810
Lieth H (1975) Modelling the primary productivity of the world. In: Lieth H, Whittaker RH (eds) Primary productivity of the biosphere. Springer, New York, pp 237–263
Linares-Palomino R, Cardona V, Hennig E, Hensen I, Hoffmann D, Lendzion J, Soto D, Herzog S, Kessler M (2009) Non-woody life-form contribution to vascular plant species richness in a tropical american forest. Plant Ecol 201:87–99
Lomolino MV (2001) Elevation gradients of species-density: historical and prospective views. Global Ecol Biogeogr 10:3–13
Mayo S, Bogner J, Boyce PC (1997) The genera of Araceae. Kew Royal Botanical Garden, London
McCain CM (2007) Could temperature and water availability drive elevational species richness patterns? A global case study for bats. Glob Ecol Biogeogr 16:1–13
McCain CM (2009) Global analysis of bird elevational diver-sity. Glob Ecol Biogeogr 18:346–360
Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J (2000) Biodiversity hotspots for conservation pri-orities. Nature 403:853–858
Ohlemu¨ller R, Wilson JB (2000) Vascular plant species richness along latitudinal and altitudinal gradients: a contribution from New Zealand temperate rainforests. Ecol Lett 3:262–266
Oksanen J, Kindt R, Legendre P, O’Hara B, Stevens MHH (2007) Vegan: Community Ecology Package. R package version 1.8-8http://r-forge.r-project.org/project/vegan/
Poulsen AD, Pendry CA (1995) Inventories of ground herbs at three different altitudes on Bukit Belalong, Brunei, Borneo. Biodivers Conserv 4:745–757
Poulsen AD, Hafashimana D, Eilu G, Innocent B, Corneille E, Ewango N, Terese BH (2005) Composition and species richness of forest plants along the Albertine Rift, Africa. Biol Skrif 55:129–143
R Development Core Team (2010) R: A language and envi-ronment for statistical computing. R Foundation for Sta-tistical Computing, Vienna
Rahbek C (1995) The elevational gradient of species richness: a uniform pattern? Ecography 18:200–205
Rahbek C (2005) The role of spatial scale and the perception of large-scale species-richness patterns. Ecol Lett 8:224–239 Ricklefs RE (2004) A comprehensive framework for global
patterns in biodiversity. Ecol Lett 7:1–15
Rosenzweig ML, Ziv Y (1999) The echo pattern of species diversity: pattern and process. Ecography 22:614–628
Sa´nchez-Gonza´lez A, Lo´pez-Mata L (2005) Plant species richness and diversity along an altitudinal gradient in the Sierra Nevada, Mexico. Divers Distrib 11:567–575 Stevens GC (1989) The latitudinal gradient in geographical
range: how so many species coexist in the tropics. Am Nat 133:240–256
The Nature Conservancy (2004) Lore Lindu National Park Management Plan 2004–2029
Tsvuura Z, Griffith ME, Gunton RM, Franks PJ, Lawes MJ (2010) Ecological filtering by a dominant herb selects for shade tolerance in the tree seedling community of coastal dune forest. Oecologia 164:861–870
Wagenmakers EJ, Farrell S (2004) AIC model selection using Akaike weights. Psychon Bull Rev 11:192–196
